arXiv subjects

Zhichen Zhao

Publications and source records attributed to Zhichen Zhao.

17 recordsLinked to original sources

arXiv2026

OneShot: Index-in-Ranking with Neural Scoring for Large-Scale Retrieval

In modern recommendation systems, retrieval serves as a primary stage responsible for filtering billions of candidate items down to thousands prior to refined ranking. To make this massive search effective and efficient, the system relies on ranking accuracy and indexing efficiency. However, these two objectives are traditionally misaligned: while the former optimizes for the alignment between ranking predictions and user behavior, the latter optimizes for a structural grouping of item representations which enables fast search among billions of candidates. Thus, despite extensive efforts to scale up interaction modeling for retrieval, they remain fundamentally limited by the structural misalignment between the ranking objectives and the proximity-learned index. In this work, we address this long-standing dichotomy by proposing a new holistic retrieval framework, OneShot. It is an end-to-end, in-model index learning framework that natively aligns index learning with ranking objectives. Using this joint learning as a structural foundation, OneShot pushes the boundaries of retrieval expressiveness by scaling interaction modeling with neural scoring beyond the persistent dot-product bottleneck. OneShot is fully deployed in Instagram's industrial short-video recommendation system, driving significant wins in user daily sessions, engagement, and time-spent. Additionally, OneShot achieves a $20\%$ recall gain at the operational ranking volume and a 10x efficiency improvement at an equivalent recall level.

cs.IR↗

arXiv2026

Efficient Inference for Incremental Causal Effects of Time to Treatment

We consider continuous time to treatment initiation. This can commonly occur in preventive medicine, such as disease screening and vaccination; it can also occur with non-fatal health conditions such as HIV infection without the onset of AIDS. While traditional causal inference focused on `when to treat' and its effects, we consider the incremental causal effect when the intensity of time to treatment initiation is intervened upon. We derive the efficient influence function for this estimand and develop an estimation framework that accommodates flexible machine learning methods while achieving fast convergence rates. Valid confidence bands are obtained leveraging empirical process theory. We illustrate our approach via simulation, and apply it to cervical cancer screening data to study the incremental effect of time to subsequent HPV testing on cervical intraepithelial neoplasia detection.

stat.ME↗

arXiv2026

Polarization Engineering of Second-Harmonic Generation in 3R-MoS$_2$ Waveguides

Chip-scale nonlinear optics enables strong light-matter interactions within compact devices, serving as a fundamental platform for multifunctional integrated photonics from classical optical signal processing to quantum information technologies. Transition metal dichalcogenide (TMDC) waveguides have recently emerged as a highly promising platform owing to their giant material nonlinearity and extended interaction lengths. To date, however, research has predominantly focused on conversion efficiency, leaving the mechanisms governing the polarization state of nonlinear signal largely unexplored. Here, we establish a comprehensive framework for engineering the polarization of second-harmonic generation (SHG) in 3R-MoS$_2$ waveguides. By synergizing polarization-resolved measurements with theoretical modeling, we reveal that the SHG polarization is determined by guided-mode interactions constrained by waveguide geometry and crystal symmetry, and further reshaped during propagation. We demonstrate that thickness-dependent guided-mode confinement and in-plane crystal symmetry provide robust, static control over SHG polarization, while propagation length offers a dynamic tuning knob for continuously tailoring the nonlinear output. Our findings provide a deterministic approach for on-chip polarization engineering, opening opportunities for reconfigurable nonlinear light sources and quantum photonic circuits.

physics.optics↗

arXiv2025

Real-time Indexing for Large-scale Recommendation by Streaming Vector Quantization Retriever

Retrievers, which form one of the most important recommendation stages, are responsible for efficiently selecting possible positive samples to the later stages under strict latency limitations. Because of this, large-scale systems always rely on approximate calculations and indexes to roughly shrink candidate scale, with a simple ranking model. Considering simple models lack the ability to produce precise predictions, most of the existing methods mainly focus on incorporating complicated ranking models. However, another fundamental problem of index effectiveness remains unresolved, which also bottlenecks complication. In this paper, we propose a novel index structure: streaming Vector Quantization model, as a new generation of retrieval paradigm. Streaming VQ attaches items with indexes in real time, granting it immediacy. Moreover, through meticulous verification of possible variants, it achieves additional benefits like index balancing and reparability, enabling it to support complicated ranking models as existing approaches. As a lightweight and implementation-friendly architecture, streaming VQ has been deployed and replaced all major retrievers in Douyin and Douyin Lite, resulting in remarkable user engagement gain.

cs.IR↗

arXiv2024

Incremental Causal Effect for Time to Treatment Initialization

We consider time to treatment initialization. This can commonly occur in preventive medicine, such as disease screening and vaccination; it can also occur with non-fatal health conditions such as HIV infection without the onset of AIDS; or in tech industry where items wait to be reviewed manually as abusive or not, etc. While traditional causal inference focused on `when to treat' and its effects, including their possible dependence on subject characteristics, we consider the incremental causal effect when the intensity of time to treatment initialization is intervened upon. We provide identification of the incremental causal effect without the commonly required positivity assumption, as well as an estimation framework using inverse probability weighting. We illustrate our approach via simulation, and apply it to a rheumatoid arthritis study to evaluate the incremental effect of time to start methotrexate on joint pain.

stat.ME↗

arXiv2024

Selective Excitation of Bloch Modes in Canalized Polaritonic Crystals

Polaritonic crystals (PoCs) have experienced significant advancements through involving hyperbolic polaritons in anisotropic materials such as $α$-MoO$_{\rm 3}$, offering a promising approach for nanoscale light control and improved light-matter interactions. Notably, twisted bilayer $α$-MoO$_{\rm 3}$ enables tunable iso-frequency contours (IFCs), especially generating flat IFCs at certain twist angles, which could enhance mode selectivity in their PoCs through the highly collimated and canalized polaritons. This study unveils the selective excitation of Bloch modes in PoCs with square-lattice structures on twisted bilayer $α$-MoO$_{\rm 3}$ with canalized phonon polaritons. Through the optimization of the square lattice design, there is an effective redistribution of canalized polaritons into the reciprocal lattices of PoCs. Fine-tuning the periodicity and orientation of the hole lattice enables momentum matching between flat IFCs and co-linear reciprocal points, allowing precise and directional control over desired Bragg resonances and Bloch modes. This research establishes a versatile platform for tunable polaritonic devices and paves the way for advanced photonic applications.

physics.optics↗

arXiv2024

Trinity: Syncretizing Multi-/Long-tail/Long-term Interests All in One

Interest modeling in recommender system has been a constant topic for improving user experience, and typical interest modeling tasks (e.g. multi-interest, long-tail interest and long-term interest) have been investigated in many existing works. However, most of them only consider one interest in isolation, while neglecting their interrelationships. In this paper, we argue that these tasks suffer from a common "interest amnesia" problem, and a solution exists to mitigate it simultaneously. We figure that long-term cues can be the cornerstone since they reveal multi-interest and clarify long-tail interest. Inspired by the observation, we propose a novel and unified framework in the retrieval stage, "Trinity", to solve interest amnesia problem and improve multiple interest modeling tasks. We construct a real-time clustering system that enables us to project items into enumerable clusters, and calculate statistical interest histograms over these clusters. Based on these histograms, Trinity recognizes underdelivered themes and remains stable when facing emerging hot topics. Trinity is more appropriate for large-scale industry scenarios because of its modest computational overheads. Its derived retrievers have been deployed on the recommender system of Douyin, significantly improving user experience and retention. We believe that such practical experience can be well generalized to other scenarios.

cs.IR↗

arXiv2021

POSO: Personalized Cold Start Modules for Large-scale Recommender Systems

Recommendation for new users, also called user cold start, has been a well-recognized challenge for online recommender systems. Most existing methods view the crux as the lack of initial data. However, in this paper, we argue that there are neglected problems: 1) New users' behaviour follows much different distributions from regular users. 2) Although personalized features are involved, heavily imbalanced samples prevent the model from balancing new/regular user distributions, as if the personalized features are overwhelmed. We name the problem as the "submergence" of personalization. To tackle this problem, we propose a novel module: Personalized COld Start MOdules (POSO). Considering from a model architecture perspective, POSO personalizes existing modules by introducing multiple user-group-specialized sub-modules. Then, it fuses their outputs by personalized gates, resulting in comprehensive representations. In such way, POSO projects imbalanced features to even modules. POSO can be flexibly integrated into many existing modules and effectively improves their performance with negligible computational overheads. The proposed method shows remarkable advantage in industrial scenario. It has been deployed on the large-scale recommender system of Kwai, and improves new user Watch Time by a large margin (+7.75%). Moreover, POSO can be further generalized to regular users, inactive users and returning users (+2%-3% on Watch Time), as well as item cold start (+3.8% on Watch Time). Its effectiveness has also been verified on public dataset (MovieLens 20M). We believe such practical experience can be well generalized to other scenarios.

cs.IR↗

arXiv2020

Short Video-based Advertisements Evaluation System: Self-Organizing Learning Approach

With the rising of short video apps, such as TikTok, Snapchat and Kwai, advertisement in short-term user-generated videos (UGVs) has become a trending form of advertising. Prediction of user behavior without specific user profile is required by advertisers, as they expect to acquire advertisement performance in advance in the scenario of cold start. Current recommender system do not take raw videos as input; additionally, most previous work of Multi-Modal Machine Learning may not deal with unconstrained videos like UGVs. In this paper, we proposed a novel end-to-end self-organizing framework for user behavior prediction. Our model is able to learn the optimal topology of neural network architecture, as well as optimal weights, through training data. We evaluate our proposed method on our in-house dataset. The experimental results reveal that our model achieves the best performance in all our experiments.

cs.MM↗

arXiv2019

Effective Domain Knowledge Transfer with Soft Fine-tuning

Convolutional neural networks require numerous data for training. Considering the difficulties in data collection and labeling in some specific tasks, existing approaches generally use models pre-trained on a large source domain (e.g. ImageNet), and then fine-tune them on these tasks. However, the datasets from source domain are simply discarded in the fine-tuning process. We argue that the source datasets could be better utilized and benefit fine-tuning. This paper firstly introduces the concept of general discrimination to describe ability of a network to distinguish untrained patterns, and then experimentally demonstrates that general discrimination could potentially enhance the total discrimination ability on target domain. Furthermore, we propose a novel and light-weighted method, namely soft fine-tuning. Unlike traditional fine-tuning which directly replaces optimization objective by a loss function on the target domain, soft fine-tuning effectively keeps general discrimination by holding the previous loss and removes it softly. By doing so, soft fine-tuning improves the robustness of the network to data bias, and meanwhile accelerates the convergence. We evaluate our approach on several visual recognition tasks. Extensive experimental results support that soft fine-tuning provides consistent improvement on all evaluated tasks, and outperforms the state-of-the-art significantly. Codes will be made available to the public.

cs.CV↗

arXiv2017

Fingerprinting molecular relaxation in deformed polymers

The flow and deformation of macromolecules is ubiquitous in nature and industry, and an understanding of this phenomenon at both macroscopic and microscopic length scales is of fundamental and practical importance. Here we present the formulation of a general mathematical framework, which could be used to extract, from scattering experiments, the molecular relaxation of deformed polymers. By combining and modestly extending several key conceptual ingredients in the literature, we show how the anisotropic single-chain structure factor can be decomposed by spherical harmonics and experimentally reconstructed from its cross sections on the scattering planes. The resulting wavenumber-dependent expansion coefficients constitute a characteristic fingerprint of the macromolecular deformation, permitting detailed examinations of polymer dynamics at the microscopic level. We apply this approach to survey a long-standing problem in polymer physics regarding the molecular relaxation in entangled polymers after a large step deformation. The classical tube theory of Doi and Edwards predicts a fast chain retraction process immediately after the deformation, followed by a slow orientation relaxation through the reptation mechanism. This chain retraction hypothesis, which is the keystone of the tube theory for macromolecular flow and deformation, was critically examined by analyzing the fine features of the two-dimensional anisotropic spectra from small-angle neutron scattering by entangled polystyrenes. It is shown that the unique scattering patterns associated with the chain retraction mechanism were not experimentally observed. This result calls for a fundamental revision of the current theoretical picture for nonlinear rheological behavior of entangled polymeric liquids.

cond-mat.soft↗

arXiv2016

Single Image Action Recognition using Semantic Body Part Actions

In this paper, we propose a novel single image action recognition algorithm which is based on the idea of semantic body part actions. Unlike existing bottom up methods, we argue that the human action is a combination of meaningful body part actions. In detail, we divide human body into five parts: head, torso, arms, hands and legs. And for each of the body parts, we define several semantic body part actions, e.g., hand holding, hand waving. These semantic body part actions are strongly related to the body actions, e.g., writing, and jogging. Based on the idea, we propose a deep neural network based system: first, body parts are localized by a Semi-FCN network. Second, for each body parts, a Part Action Res-Net is used to predict semantic body part actions. And finally, we use SVM to fuse the body part actions and predict the entire body action. Experiments on two dataset: PASCAL VOC 2012 and Stanford-40 report mAP improvement from the state-of-the-art by 3.8% and 2.6% respectively.

cs.CV↗

arXiv2012

Holographic Lifshitz flows and the null energy condition

We study asymptotically Lifshitz spacetimes and the constraints on flows between Lifshitz fixed points imposed by the null energy condition. In contrast with the relativistic holographic c-theorem, where the effective AdS radius, L, is monotonically decreasing in the IR, for Lifshitz backgrounds we find that both L and z may flow in either direction. We demonstrate this with several numerical examples in a phenomenological model with a massive gauge field coupled to a real scalar.

hep-th↗

arXiv2012

Holographic c-theorems and higher derivative gravity

In AdS/CFT, the holographic Weyl anomaly computation relates the a-anomaly coefficient to the properties of the bulk action at the UV fixed point. This universal behavior suggests the possibility of a holographic c-theorem for the a-anomaly under flows to the IR. We prove such a c-theorem for higher curvature Lovelock gravity, where the bulk equations of motion remain second order. We also explore f(R) gravity as a toy model where higher derivatives cannot be avoided. In this case, monoticity of the flow requires an additional condition related to the higher derivative nature of the theory. This is in contrast to the case of f(R) black hole entropy, where the second law follows from application of the full Einstein equations and the null energy condition.

hep-th↗

arXiv2011

A holographic c-theorem for higher derivative gravity

We prove a holographic c-theorem for the a central charge in AdS/CFT where the bulk is described by a gravitational action built out of an arbitrary function f(R^{ab}_{cd}) of the Riemann tensor coupled to bulk matter. This theorem holds provided a generalized null energy condition involving both matter and higher curvature gravitational interactions is satisfied. As an example, we consider the case of a curvature-squared action, and find that generically the generalized null energy condition involves not just the bulk matter, but also the sign of R" where a prime denotes a radial derivative and where R is the bulk scalar curvature.

hep-th↗

arXiv2011

Supersymmetric massive truncations of IIb supergravity on Sasaki-Einstein manifolds

Motivated by recent interest in applications of the AdS/CFT correspondence to condensed matter applications involving fermions, we present the supersymmetric completion of the recent massive truncations of IIB supergravity on Sasaki-Einstein manifolds. In particular, we reduce the fermionic sector of IIB supergravity to obtain five dimensional N=2 supergravity coupled to one hypermultiplet and one massive vector multiplet. The supersymmetry transformations and equations of motion are presented and analyzed. Finally, a particularly interesting truncation to N=2 supergravity coupled to a single hypermultiplet is presented which is the supersymmetric completion of the recently constructed bosonic theory dual to a 3+1 dimensional system exhibiting a superconducting phase transition.

hep-th↗

arXiv2010

Consistent massive truncations of IIB supergravity on Sasaki-Einstein manifolds

Recent work on holographic superconductivity and gravitational duals of systems with non-relativistic conformal symmetry have made use of consistent truncations of D=10 and D=11 supergravity retaining some massive modes in the Kaluza-Klein tower. In this paper we focus on reductions of IIB supergravity to five dimensions on a Sasaki-Einstein manifold, and extend these previous truncations to encompass the entire bosonic sector of gauged D=5, N=2 supergravity coupled to massive multiplets up to the second Kaluza-Klein level. We conjecture that a necessary condition for the consistency of massive truncations is to only retain the lowest modes in the massive trajectories of the Kaluza-Klein mode decomposition of the original fields. This is an extension of the well-known result that consistent truncations may be obtained by restricting to the singlet sector of the internal symmetry group.

hep-th↗